Hsp104 is a protein chaperone that helps cells recover from stress by resolubilizing proteins from aggregates. This disaggregation activity depends on assistance by Hsp40 and Hsp70 and is required for propagation of the amyloid-based yeast prions. Fifteen years ago it was discovered that when Hsp104 expression is elevated under conditions where it is not normally induced it causes PSI prions to be lost from dividing cells. A universally held view of the underlying mechanism is that Hsp104 protein disaggregation activity dissolves prion aggregates until templates for further prion propagation are eliminated. However, most prions are unaffected by elevating Hsp104, suggesting that PSI is hypersensitive to this disaggregation or that something else is happening. We identified mutants that showed Hsp104's amino-terminal domain (NTD), which is conserved but whose function is unknown, is required for prion elimination by overproduced Hsp104 but not for Hsp104 activities required for thermotolerance, protein disaggregation and prion propagation. Excess NTD alone does not cure prions. These findings imply that Hsp104's known functions are insufficient for eliminating prions. Thus, the mechanism of prion curing does not appear to be caused by complete solubilization of prion aggregates but by an unknown Hsp104 function. The curing-defective Hsp104 mutants also suppress anti-prion effects of a mutant Hsp70, which suggested that other alterations that suppress this mutant Hsp70 will also interfere with Hsp104 curing of PSI. We found deleting Sti1p, a co-chaperone regulator of Hsp70 that we previously showed acted as such a suppressor, disrupted Hsp104's prion curing activity. In addition to Hsp40 and Hsp70, Sti1p is part of the Hsp90 chaperone machinery, which acts in a folding maturation pathway for many "client" proteins, such as signaling and transcription factors. We showed the role of Sti1p in Hsp104 curing depended on its ability to regultate both Hsp70 and Hsp90, and that the region of Hsp104 that mediates physical interaction with Sti1p was dispensable for curing, indicating the effects were indirect. We further showed that certain alterations in Hsp90 also inhibited the curing. Together our data suggest the role of Sti1p in curing by elevated Hsp104 is indirect through regulation of the Hsp90 machinery.